The adrenal hormones dehydroepiandrosterone (DHEA) and its sulfate (DHEA-S) have been negatively correlated with insulin levels, cardiovascular morbidity, cancer, obesity and diabetes. However, the mechanism of the association is unclear. Insulin acts as a physiological regulator of DHEA/DHEA-S metabolism; hyperinsulinemia decreases serum DHEA-S in normal subjects and in women with polycystic ovary syndrome (PCOS), while increasing androstenedione (D4-A) and testosterone levels conversely. The divergent correlations of circulating DHEA-S and testosterone/adrostenedione combined with the animal data of diet effects of DHEA suggest that DHEA-S and testosterone may therefore exert opposing effects on insulin secretion and action. Women with the Polycystic Ovary Syndrome (PCOS) have irregular menstrual cycles and hyperandogenism. In addition, women with PCOS have been shown to be insulin resistant, independent of their frequent obesity. Women with PCOS also often have elevated mean levels of LH, with an increased LH pulse frequency and amplitude. The increases in both pulse amplitude, which could be secondary to an increased pituitary responsiveness and/or an increased amount of GnRH stimulating the pituitary, and in pulse frequency, which could only be secondary to an increase in hypothalamic GnRH frequency, suggests a hypothalamic site of action and possibly also a pituitary site of action as well. Insulin and LH have both been shown to increase ovarian androgen production in vitro and elevations of either may contribute to excess ovarian androgen production in vivo. Insulin also lowers sex hormone binding globulin levels (SHBG) thereby increasing free androgen concentrations. Finally, it is possible that hyperinsulinemia contributes to the elevated LH levels as there is in vitro evidence that the pituitary has insulin receptors and that insulin may directly modulate pituitary LH release. Thus, this study is designed to examine the role of hyperinsulinemia in elevated LH levels in women. Data analysis will be performed at study end. To date, 11 of the 40 anticipated subjects have been enrolled.